Nesne Tespiti
The task of computerized vision that determines both class and location of objects in a image.
Object detection (object detection) is the task of a computerized view, which belongs to the class of one or more objects in a image, as well as precisely where it is located in the image. The model usually draws a limiting box (bounding box) around each object detected and throws a trust score with a class tag to this box. This is a much more complex task than image classification that only responds to the question of "Is there a cat in this image" because the model should simultaneously distinguish multiple objects in different sizes and locations.
The architectures such as ROADO (You Only Look Once), Faster R-CNN and SSD have been important milestones in object detection, especially ROADO is known to achieve real-time speed by processing the image in one pass. Object detection is the basis of many critical applications such as recognition of autonomous vehicles, other tools and traffic signs, monitoring suspicious behaviors of security cameras, inventory tracking in retail stores and marking abnormal tissues in medical images. The accuracy of the model is usually measured with metrics such as IoU (Intersection over Union) and mAP (mean Average Precision).
The architectures such as ROADO (You Only Look Once), Faster R-CNN and SSD have been important milestones in object detection, especially ROADO is known to achieve real-time speed by processing the image in one pass. Object detection is the basis of many critical applications such as recognition of autonomous vehicles, other tools and traffic signs, monitoring suspicious behaviors of security cameras, inventory tracking in retail stores and marking abnormal tissues in medical images. The accuracy of the model is usually measured with metrics such as IoU (Intersection over Union) and mAP (mean Average Precision).